H5N1 will move to human populations eventually. There is really no getting around it.
As with all influenza viruses it is a negative-sense ssRNA virus. The single stranded RNA mutates very easily because there is no complimentary sequence to stabilize the genetic information. If it had a double strand nature it would require the cell to use an enzyme protein to separate the strands which also verifies the sequence. The negative-sense architecture means the RNA is like a photo negative of the actual RNA required for transcription and protein production. The cell will use proteins to build a mirror image RNA for transcription which seems like an extra step (it is) but instead of directly transcribing the original RNA strand one nucleotide at a time, the cell can make multiple copies of the mirror image all being transcribed together generating a huge amount of proteins very quickly for new viral synthesis.
There have been cases of human H5N1 infection. The latest CDC reports from 2007 indicate 86 cases of H5N1 Type A Influenza with 59 deaths. The infections, as with others in the past, have come from direct contact with dead poultry. As it stands now, there seems to be no human to human vector. This will change. That's the thing about probability, if it can happen, it will happen. The only thing stopping it right now is time. The rate at which ssRNA mutates is quite high. We can only hope that when it does jump to humans that the mutations also temper the the mortality.
An aggressive highly pathogenic H5N1 with human to human transmission will be devastating. Unlike Ebola, H5N1 has a longer incubation period allowing far more people to be infected before killing the host. That was the limiting factor with Ebola. It killed the victim so fast they did not have time to infect many people before dying.